Explore every episode of the podcast Joe Carlsmith Audio
| Title | Pub. Date | Duration | |
|---|---|---|---|
| On restraining AI development for the sake of safety | 19 mars 2026 | 01:22:21 | |
My take on slowing down AI. Text version here: https://joecarlsmith.com/2026/03/19/on-restraining-ai-development-for-the-sake-of-safety/ | |||
| Building AIs that do human-like philosophy | 29 janv. 2026 | 00:33:33 | |
AIs will face philosophical questions humans can't answer for them. Text version here: https://joecarlsmith.com/2026/01/29/building-ais-that-do-human-like-philosophy/ | |||
| How human-like do safe AI motivations need to be? | 12 nov. 2025 | 01:23:32 | |
AIs with alien motivations can still follow instructions safely on the inputs that matter. Text version here: https://joecarlsmith.com/2025/11/12/how-human-like-do-safe-ai-motivations-need-to-be/ | |||
| Leaving Open Philanthropy, going to Anthropic | 03 nov. 2025 | 00:32:09 | |
On a career move, and on AI-safety-focused people working at AI companies. Text version here: https://joecarlsmith.com/2025/11/03/leaving-open-philanthropy-going-to-anthropic/ | |||
| Controlling the options AIs can pursue | 29 sept. 2025 | 00:55:34 | |
On boxing AIs, and on making deals with them. Text version here: https://joecarlsmith.com/2025/09/29/controlling-the-options-ais-can-pursue | |||
| Giving AIs safe motivations | 18 août 2025 | 01:23:25 | |
A four-step picture. Text version here: https://joecarlsmith.com/2025/08/18/giving-ais-safe-motivations | |||
| The stakes of AI moral status | 21 mai 2025 | 00:37:29 | |
On seeing and not seeing souls. Text version here: https://joecarlsmith.com/2025/05/21/the-stakes-of-ai-moral-status/ | |||
| Can we safely automate alignment research? | 30 avr. 2025 | 01:29:38 | |
It's really important; we've got a real shot; there are a ton of ways to fail. Text version here: https://joecarlsmith.com/2025/04/30/can-we-safely-automate-alignment-research/. There's also a video and transcript of a talk I gave on this topic here: https://joecarlsmith.com/2025/04/30/video-and-transcript-of-talk-on-automating-alignment-research/ | |||
| AI for AI safety | 14 mars 2025 | 00:27:51 | |
We should try extremely hard to use AI labor to help address the alignment problem. Text version here: https://joecarlsmith.com/2025/03/14/ai-for-ai-safety | |||
| Paths and waystations in AI safety | 11 mars 2025 | 00:18:07 | |
On the structure of the path to safe superintelligence, and some possible milestones along the way. Text version here: https://joecarlsmith.substack.com/p/paths-and-waystations-in-ai-safety | |||
| When should we worry about AI power-seeking? | 19 févr. 2025 | 00:46:54 | |
Examining the conditions required for rogue AI behavior. Text version here: https://joecarlsmith.substack.com/p/when-should-we-worry-about-ai-power | |||
| What is it to solve the alignment problem? | 13 févr. 2025 | 00:40:13 | |
Also: to avoid it? Handle it? Solve it forever? Solve it completely? Text version here: https://joecarlsmith.substack.com/p/what-is-it-to-solve-the-alignment | |||
| How do we solve the alignment problem? | 13 févr. 2025 | 00:08:43 | |
Introduction to a series of essays about paths to safe and useful superintelligence. Text version here: https://joecarlsmith.substack.com/p/how-do-we-solve-the-alignment-problem | |||
| Fake thinking and real thinking | 28 janv. 2025 | 01:18:47 | |
When the line pulls at your hand. | |||
| Takes on "Alignment Faking in Large Language Models" | 18 déc. 2024 | 01:27:54 | |
What can we learn from recent empirical demonstrations of scheming in frontier models? Text version here: https://joecarlsmith.com/2024/12/18/takes-on-alignment-faking-in-large-language-models/ | |||
| (Part 2, AI takeover) Extended audio from my conversation with Dwarkesh Patel | 30 sept. 2024 | 02:07:33 | |
Extended audio from my conversation with Dwarkesh Patel. This part focuses on the basic story about AI takeover. Transcript available on my website here: https://joecarlsmith.com/2024/09/30/part-2-ai-takeover-extended-audio-transcript-from-my-conversation-with-dwarkesh-patel | |||
| (Part 1, Otherness) Extended audio from my conversation with Dwarkesh Patel | 30 sept. 2024 | 03:58:38 | |
Extended audio from my conversation with Dwarkesh Patel. This part focuses on my series "Otherness and control in the age of AGI." Transcript available on my website here: https://joecarlsmith.com/2024/09/30/part-1-otherness-extended-audio-transcript-from-my-conversation-with-dwarkesh-patel/ | |||
| Introduction and summary for "Otherness and control in the age of AGI" | 21 juin 2024 | 00:12:23 | |
This is the introduction and summary for my series "Otherness and control in the age of AGI." | |||
| Second half of full audio for "Otherness and control in the age of AGI" | 18 juin 2024 | 04:11:02 | |
Second half of the full audio for my series on how agents with different values should relate to one another, and on the ethics of seeking and sharing power. | |||
| First half of full audio for "Otherness and control in the age of AGI" | 17 juin 2024 | 03:07:29 | |
First half of the full audio for my series on how agents with different values should relate to one another, and on the ethics of seeking and sharing power. | |||
| Loving a world you don't trust | 17 juin 2024 | 01:03:54 | |
Garden, campfire, healing water. | |||
| On attunement | 25 mars 2024 | 00:44:14 | |
Examining a certain kind of meaning-laden receptivity to the world. | |||
| On green | 21 mars 2024 | 01:15:13 | |
Examining a philosophical vibe that I think contrasts in interesting ways with "deep atheism." | |||
| On the abolition of man | 18 janv. 2024 | 01:09:22 | |
What does it take to avoid tyranny towards to the future? | |||
| Being nicer than Clippy | 16 janv. 2024 | 00:47:30 | |
Let's be the sort of species that aliens wouldn't fear the way we fear paperclippers. | |||
| An even deeper atheism | 11 janv. 2024 | 00:25:12 | |
Who isn't a paperclipper? | |||
| Does AI risk "other" the AIs? | 09 janv. 2024 | 00:13:15 | |
Examining Robin Hanson's critique of the AI risk discourse. | |||
| When "yang" goes wrong | 08 janv. 2024 | 00:21:32 | |
On the connection between deep atheism and seeking control. | |||
| Deep atheism and AI risk | 04 janv. 2024 | 00:46:59 | |
On a certain kind of fundamental mistrust towards Nature. | |||
| Gentleness and the artificial Other | 02 janv. 2024 | 00:22:39 | |
AIs as fellow creatures. And on getting eaten. | |||
| In search of benevolence (or: what should you get Clippy for Christmas?) | 27 déc. 2023 | 00:52:52 | |
What is altruism towards a paperclipper? Can you paint with all the colors of the wind at once? | |||
| The goal-guarding hypothesis (Section 2.3.1.1 of "Scheming AIs") | 16 nov. 2023 | 00:19:11 | |
This is section 2.3.1.1 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| Summing up "Scheming AIs" (Section 5) | 16 nov. 2023 | 00:15:46 | |
This is section 5 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| How useful for alignment-relevant work are AIs with short-term goals? (Section 2.2.4.3 of "Scheming AIs") | 16 nov. 2023 | 00:09:21 | |
This is section 2.2.4.3 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| Is scheming more likely if you train models to have long-term goals? (Sections 2.2.4.1-2.2.4.2 of "Scheming AIs") | 16 nov. 2023 | 00:09:01 | |
This is sections 2.2.4.1-2.2.4.2 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| Speed arguments against scheming (Section 4.4-4.7 of "Scheming AIs") | 16 nov. 2023 | 00:15:19 | |
This is section 4.4 through 4.7 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| Simplicity arguments for scheming (Section 4.3 of "Scheming AIs") | 16 nov. 2023 | 00:19:37 | |
This is section 4.3 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| The counting argument for scheming (Sections 4.1 and 4.2 of "Scheming AIs") | 16 nov. 2023 | 00:10:40 | |
This is sections 4.1 and 4.2 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| Arguments for/against scheming that focus on the path SGD takes (Section 3 of "Scheming AIs") | 16 nov. 2023 | 00:29:03 | |
This is section 3 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| Non-classic stories about scheming (Section 2.3.2 of "Scheming AIs") | 16 nov. 2023 | 00:24:34 | |
This is section 2.3.2 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| Empirical work that might shed light on scheming (Section 6 of "Scheming AIs") | 16 nov. 2023 | 00:28:00 | |
This is section 6 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| Does scheming lead to adequate future empowerment? (Section 2.3.1.2 of "Scheming AIs") | 16 nov. 2023 | 00:22:54 | |
This is section 2.3.1.2 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| "Clean" vs. "messy" goal-directedness (Section 2.2.3 of "Scheming AIs") | 16 nov. 2023 | 00:16:44 | |
This is section 2.2.3 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| Two sources of beyond-episode goals (Section 2.2.2 of "Scheming AIs") | 16 nov. 2023 | 00:21:25 | |
This is section 2.2.2 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| Two concepts of an "episode" (Section 2.2.1 of "Scheming AIs") | 16 nov. 2023 | 00:12:08 | |
This is section 2.2.1 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| Situational awareness (Section 2.1 of "Scheming AIs") | 16 nov. 2023 | 00:09:27 | |
This is section 2.1 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| On "slack" in training (Section 1.5 of "Scheming AIs") | 16 nov. 2023 | 00:07:12 | |
This is section 1.5 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| Why focus on schemers in particular? (Sections 1.3-1.4 of "Scheming AIs") | 16 nov. 2023 | 00:31:17 | |
This is sections 1.3-1.4 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| A taxonomy of non-schemer models (Section 1.2 of "Scheming AIs") | 16 nov. 2023 | 00:11:20 | |
This is section 1.2 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||
| Varieties of fake alignment (Section 1.1 of "Scheming AIs") | 16 nov. 2023 | 00:17:54 | |
This is section 1.1 of my report “Scheming AIs: Will AIs fake alignment during training in order to get power?” Text of the report here: https://arxiv.org/abs/2311.08379 | |||